An automated, high-throughput sequence read classification pipeline for preliminary genome characterization.

In the absence of a complete genome sequence, considerable insight into genome structure can be gained from survey sequencing of genomic DNA. To facilitate high-throughput characterization of genome structure based on shotgun sequence reads, we have developed an automated sequence read classification pipeline (SRCP). The SRCP uses a battery of novel and standard sequence analysis algorithms along with a sophisticated decision tree to place reads into "best fit" functional/descriptive categories. Once "primed" with genomic sequence data, the SRCP also permits estimation of gene/repeat enrichment afforded by reduced-representation sequencing techniques. To our knowledge, the SRCP is the only tool that has been designed to provide a description of a genome or a genome component based on sample sequence reads. In an initial test of the SRCP using sequence data from Sorghum bicolor, it was shown to provide results similar in quality to results generated by manual classification. Although the SRCP is not a replacement for manual sequence characterization, it can provide a rapid, high-quality overview of genome sequence content and facilitate subsequent annotation. The SRCP presumably can be adapted for analysis of any eukaryotic genome.